Thixotropic Behavior of Oils RUTH N. WELTMANN Interchemical Corporation, New York, N. Y
Various types of oils in the viscosity range of 1 to 800 poises were measured on a rotational viscometer capable of imparting shearing stresses over a wide range. From these oils, flow curves were obtained extending from low to high rates of shear. All these oils showed a definite thixotropic behavior and exhibited all the characteristics of thixotropic plastics above a certain
rate of shear designated as “the limiting rate of shear”. Below this critical point the oils behaved like true Newtonian liquids, showing no signs of thixotropic structure. The limiting rates of shear were found to be related to the measured true Newtonian viscosities of the oils. The product of limiting rate of shear and viscosity was a constant for all the oils tested.
0
tion of sufficiently high rates of shear, or only onepoint visILS have been generally assumed to be true Newtonian cosity determinations have usually been performed, and liquids, even t o the extent that industrial laboratories are these are insufficient to show thixotropy. The rotational satisfied to determine the viscosity of an oil by a one-point viscometer developed in this laboratory, described by Green measurement. It has been suggested by some workers in the (S), was well suited for studying the thixotropic behavior field that certain oils exhibit a pseudoplastic characteristic. of oils, Extensive studies in this laboratory have shown that the many different oils investiInstrument gated are not pseudoplastics The viscometer has played an but are true Newtonian liquids important part in obtaining the reI sults reported below. It is built on for a limited range of rate of the rotational principle, where the shear and behave, beyond the cup is rotated at various speeds and limited range, at least for the the bob is stationary, being suspended heavier oils, like thixotropic from a helical spring. The torsional modulus of this spring is calibrated plastics. by weights. Various springe are used I n the literature (2, 9) thixotto cover a wider range of viscosity ropy is defined as an isothermal measurements, The dimensions of gel-sol-gel transformation. Thixothe cup and the bob have been chosen to minimize the effect of plug tropic behavior does not require a flow and turbulence. A lid on the complete transformation; it is concup prevents the oils from climbing sidered a sufficient condition for up the shaft of the bob and from the presence of thixotropy that a being thrown out of the cup at higher rates of shear. No “channelin material changes its plastic viscosity could be evidenced, provided from a higher to a lower value as a bob was perfectly in center with the result of mechanical agitation and cup. The end effect introduced by regains its original high viscosity the bottom of the bob is about 2 per cent of the whole shearing effect upon rest. However, the thixoand therefore can be disregarded in tropic characteristic of a material calculatingviscositiesand yieldvalues. is also a function of time, which A constant temperature bath keeps means that the viscosity of the the temperature of the material to be investigated within =!=0.2”C. thixotropic material depends not only on previous mechanical agitaFlow Curves tion but also on the time period during which the material has been The viscosities and the yield subjected to such mechanical agitavalues of the different materials tion. are obtained from flow curves, found by plotting the number of Thixotropy is found in paints (8, revolutions per minute as a function 10,11), where it often proves to be of the resulting torque. In the useful, in bentonite suspensions (7), process of getting such flow curves, in gelatin sols (8),in iron oxide the speed of the cup is changed sols (IS), and in printing inks ( 3 ) . and the corresponding deflection Little or no reference has been made of the bob, a measure of the to the thixotropy of oils. This FIQURE1. FLOWCURVES torque, is marked down. It is thixotropic characteristic of oils Upper. A . Nonthixotropic true Newtonian understood that the amount of deprobably has not been recogflection depends upon nized because most instruments the angular velocity of the cup. used for viscosity measurements Since the angular velocities or have not permitted the applica-
tfi
424
ANALYTICAL EDITION
July 15, 1943
the revolutions per minute are proportional to the rate of shear or to the velocity gradient, in the flow curve diagram, the revolutions per minute of the ordinate may be replaced by a rate of shear ordinate. The relationship between the revolutions per minute (r. p. m.) and velocity gradient or rate of shear (dv/dr) in set.-" for a rotational instrument is 1 r. p. m. = 60 r2hSdv/dr
(1)
where P is any radius between the cup and the bob, h is the immersed height of the bob, and S is an instrumental constant equal to (1/R:
velopment of some curvature, explained (1, 4, 18) as caused by plug flow. Experimental data have been given by Green (3). However, the curvature in most cases where pigment suspehsions and oils were investigated shows a greater extension than would be expected from plug flow. This extended curvature at lower rates of shear is believed to be caused mostly by thixot,ropy. The plastic viscosities and the yield values relating to the straight portion of the downcurve may be obtained from Reiner’s equation as follows: (4)
- l/R;)/4sh
f = T2C
where R , is the radius of the cup and Rb is the radius of the bob.
%i.
(5)
where U is the plastic viscosity in poises and T is the torsion in dynes-centimeter. T 2is the torsion corresponding t o the intercept which is obtained by extending the straight portion of the flow curve to the torsion axis, w is the angular velocity, f is the yield value in dynes per s uare centimeter, and C is an instrumental constant equal to ‘i/ln(R,/Rb).
A (Figure 1, center) represents a flow curve obtained from a nonthixotropic true plastic, and it is characterized by the fact that the up- and downcurves coincide. The criterion of any plastic is the presence of an intercept of the downcurve with the torsion axis, indicating the existence of yield value. B (center) is a flow curve obtained from a thixotropic true plastic, where the upcurve has a curvature throughout all rates of shear and does not coincide with the downcurve. Finally (Figure 1, lower) two flow curves are shown, obtained from pseudoplastic materials. A represents a nonthixotropic and B thixotropic material. Such pseudoplastic flow curves are characterized by the fact that even their downcurves have a curvature throughout all speeds.
K
I
425
/ PLASTIC VISCOSITY U
FIGURE 2. SCHEMATIC CURVE SHOWING CHANGE IN PLASTICVISCOSITY WITH TOPR. P. M. FOR THIXOTROPIC PLASTIC
Thixotropy of Oils The mean rate of shear can be calculated for a given r. p. m. by substituting a mean radius in Equation 1. Then the equation for the mean rate of shear expressed in reciprocal seconds is dv/dr = 4 r. p. m./60 hS ( R ,
+ Ra)Z
(2)
Though the measurements were made with cups and bobs of various sizes, the reported revolutions per minute weie recalculated for a cup of 1.5-cm. radius and a bob of 1.3-cm. radius and 5.1-cm. immersed height. Then, the relationship between the revolutions per minute and the mean velocity gradient or the mean rate of shear (dvldr) in sec.-I is 1 r. p. m. = 1.36 (du/dr)
Any thixotropic plastic mill show a viscosity which depends upon the highest rate of shear to which the material has been subjected before starting on the downcurve. These downcurves are always found to be straight lines, thus indicating a stable condition. This stable condition has been termed thixotropic level by Green (3) and can be identified by its top rate of shear or top r. p. m.
(3)
The torque, also called the shearing stress, may be expressed in dynes-centimeter, a value which can be obtained by multiplying the deflection by the torsional constant of the helical spring. Flow curves are obtained by increasing the rates of shear to any desired maximum value and then decreasing them until the starting value is reached. Following this procedure hysteresis loops (3, 6, 8, 10) are obtained for thixotropic materials, while in nonthixotropic materials the up- and downcurves coincide. In Figure 1 six typical flow curves are shown. A (upper) is representative of a nonthixotropic true Newtonian liquid; the up- and downcurves coincide and form a straight line passing through the point of origin. B (upper) is representative of a thixotropic liquid; its upcurve has a continuous curvature, while its downcurve is again a straight line passing through the point of origin. A and B (center) are comparable to A and B (upper), but are obtained from true plastic materials and are representative of such substances. The downcurve of a true plastic material has a large linear portion, but a t lower rates of shear shows a d e
FIGURE 3. FLOWCURVES Hysteresis loops obtained for thixotropic plastic measured to top r. p. m. of ( A ) 100, ( B ) 200
The correlation found between the plastic viscosity and the respective top rate of shear or top r. p. m. is shown in Figure 2, drawn schematically according to an equation given in another paper (5):
U
=
In ( K / r .p. m.”/rn
(6)
where 2/m, also designated as M , is the coefficient of thixotropic breakdown and is defined as the loss in shearing force
426
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
Vol. 15, No. 7
investigated in this laboratory is their apparent normal behavior (true Newtonian) up to a certain rate of shear. Only above this (Temperature constant, 30’ C.) rate of shear, which will be called the “limitViscosity X Limiting ing rate of shear”, does the thixotropic b e Viscosity Revolutions Limiting l/Limiting Limiting (True per Minute Rates of Rates of Rates of havior of oils become apparent. This limiting Oils Newtonian) ( R P M ) Shear Shear Shear rate of shear is not a fixed value but dePoises See. -1 Sec. Dynes/sq. cm. pends essentially upon the viscosity of the oil Mineral oil0 780 12 9.0 0.113 6900 Isobutylene oil 770 12 9.0 0.113 6800 measured in its true Newtonian region. These Linseed oil (low acid) 380 25 18.4 0.054 7000 limiting rates of shear, even for highly viscous Linseed oil 250 38 28.0 0.036 7000 Isobutylene oil 180 52 38.3 0.026 6800 oils, are so high that most st.andard methods Linseed oil 125 75 55.5 0.018 6800 Linseed oil 115 81 59.8 0.017 6800 of measuring viscosities of oils did not permit Esso lubricant 3000 103 95 70.0 0,014 7200 using sufficiently high rates of shear to detect Mineral oil and g u m varnish 72 135 100 0.010 7 100 thixotropy. These limiting rates of shear are Linseed oil 48 205 151 0.007 7200 Mineral oil 31 310 228 0,004 7100 not very sharply defined, but approximate Mineral oil oil 30 315 231 0.004 6900 values are given in Table I for various Isobutylene 23 405 298 0.003 6800 7000 oils. Figure 4 shows the reciprocal of the Mineral oil 21 455 335 0.003 Mineral oila 20 490 360 0,003 7200 limiting rates of shear plotted as a function Mineral oil 19 500 370 0.003 7000 Linseed oil 16 580 430 0.002 6800 *. of their true Newtonian viscosities. This rela10 725 532 ... Linseed oil b Castor oilb 4 1450 1070 ... ...* tionship is linear, indicated by the straight Mineral oil b 3 1450 1070 ... line. Isobutylene oil6 1 1450 1070 ... .. Mineral oil (medical) b 1 1450 1070 ... .. This indicates that only a t extremely large a Standard viscosity oils supplied by National Bureau of Standards as refined mineral oil. finite values of the viscosity will the limiting b Since, owing to instrument limitations, limiting rates of shear could not be reached, maximum R P M and rates of shear t o which oils could be measured are tabulated. At rate of shear approach zero. On the other these rates of shear, below limiting rates of shear, oils still behave like true Nenrtonian hand, oils of extremely low viscosity, below 1 liquids. poise, may have a very high value of limiting rate of shear-indeed, a value so high that it may approach infinity. Though it has not been possible to determine the limiting rate of shear value for oils of low Newtonian viscosities because of limitations imposed by the viscometer, it may very well be bhat all oils are thixotropic. The force (torque) acting between t w o adjacent layers of the oil is equal to the product of viscosity and rate of shear. This product for the limiting rate of shear was found to be approximately constant for all oils (Table I) if the measurements were performed by increasing the rates of shear (or r. p. m.) very fast and by immediately decreasing them without waiting a t the top r. p. m. The shearing force acting between TABLE I. LIMITINGRATESOF SHEAR FOR OILS OF VARYING VISCOSITY (FIGURE 4)
100
50
0
FIGURE4. CH.4XGE O F RECIPROCALS O F LIMITING RATEOF SHEARWITH TRUENEWTOXIAN VISCOSITIES OF OILS
-
20
I
N
I a. IO a
-
\
per unit area per unit increase in velocity gradient, and K is an integration constant and is constant for each material. Khen flow properties of thixotropic materials are investigated, a relationship is found between the intercept on the torque axis and the respective top rate of shear. In most cases the intercept increases with an increase in top rate of shear or top r. p. m., which is shownin Figure 3 schematically. This subject will be treated in more detail by H. Green and the author in a subsequent paper. The flow curves obtained from the oils seem to be a combination between A (Figure 1, upper) and B (Figure 1, center). The outstanding fact concerning the oils which have been
5
-~
_ i
2
I
PLASTIC VISCOSITY
U (poises)
FIGURE5 . EXPERIMENTAL CURVE SHOWING CHANGEOF PLASTIC VISCOSITYWITH CHANGEIN TOPR. P. M.
ANALYTICAL EDITION
July 15, 1943
100
TABLE 11. VISCO~ITIES OF HEAVY MINERAL OIL Top Rates of Shear Sec. - 1
A.
Top R. P. M.
Plastic Viscosity
Torque Intercept Dynes cm.
Poises
Plastic viscosities and intercepts a t various thixotropic perature constant, 30" C. 112 665 s3 150 205 538 222 302 402 290 393 316 365 495 250 440 595 197 516 700 150
levels. 1.0
6.7 14.5 20.5 24.0 28.0 31.0
x
50
Tem105
7 20 0X 'J,
-a
I IO
B. True Kewtonian viscositks a t various temperatures. Top r . p. rn. below
K
limiting R P M "
Temperature O
c.
25 30 40 50 60 80 90
True Newtonian Viscosity Poises 1130 780 290 130 60 15 10
417
,
-
5
' 7i
1 I
I
I
I
i 1I
1 1
I
1 I
I
i
I I I I1 IO 20 30 40 TORQUE INTERCEPT x I U S (dynes cm.)
FIGURE 6. EXPERIJIEXTAL CURVESHOWtwo adjacent layers of the oil csblculated from the average ING CHANGE OF INTERCEPT WITH CHANGE value of the product of viscosity and limiting rate of shear IN TOPR. P. hl. was found to- be about 7000 dynes per square centimeter. This minimum shearing force of 7000 dynes per square centimeter may be required to overcome ?\IIXERAL OIL FLOW CURVES ME.4SURED TO v.\RIOVS TABLE 1x1. HEAVY an energy barrier before the alignment TOPR. P. 31. (FIGURE 7) of the micelles in the direction of rate of (Ternp-raturp constant, 30" C.) shear can start. Downcurve Torque X 10-6 f o r Various Thixotropic Levels Lpcurve Little has yet been said about the flow 112 302 405 205 393 59; 700 Torque r . p , 111. x 10-3 1 . p. 111 r. p . m. I . p. m. r. p , m. r. p. ni r p. ni. curves of oils above the limiting rates of Dynes Dynes shear. In this region these flow curves CWI. R. p . m. cm. are very much like flow curves obtained 3,% 3.0 4 6 4.25 3 36 2.48 2.3 14.5 4.95 4.6 6 5 G O 5 : 3.9 3.5 22 from true plastic materials; in fact,, an 11 9., 8.1 10.6 9.0 6.5 7.2 36 10.6 ... intercept can be found for the down18.8 16.6 ... 18.0 63.5 18 4 16:6 1314 . . .. . . . 18.6 . . . 76 curves of the hysteresis loops of oils. 25.8 23.3 ... 17.4 .. 24.8 S5 25.8 These do\\ ncurves contain fairly large 24:9 .. 27.6 .. 29.9 ,.. 31.9 112 34 straight-line portions very similar to the 27.2 ... 04.2 36.4 .. .. 36.6 b9,9 130 36.1 32:9 . . . . . . . . . 145 straight-line portions of downcurves of true .. .. .. 44.2 41.5 ... ... 158 46:0 plastics. 39.7 43.9 .. 47 35.9 180 52.3 50 Measurements furthermore indicate that ... .. 33:9 .. ... ... .. .. .. 192 ... .. 51.9 .. ... 205 56:O thixotropic levels exist which control the 46:O 52.0 .. ... ... ... 218 .. 56.7 .. ... 43.5 .. 234 60:9 ... plastic viscosity of the oil as a function 01 the top rate of shear or top r. p. m. 53.0 42.4 61 .. .. ... 57 ... "32 64 64.9 ... 52 .. ... 278 67.5 Figure 5 and Table 11, A, show the reln57:O .. .. ... 62.5 ... .. ... 289 . . . . . . . . . . . . . . . . 302 i 0 tionship of the plastic viscosity of a typi.. 51.5 .. . . . 66 312 .. . . . cal oil to the respective top r. p. m. ,4ny ... ... 61 57 .. 326 73 ... doubt regarding the plasticity of oils above .. .. ... ... ... .. ... 349 74.8 ... ... 65 ... .. .. ... .. 360 the limiting rates of shear is removed by 56.5 ... 74.0 61 ... 370 76 the similarity of Figures 2 and 5. 69:l ... ... .. ... ... 393 77.8 But there is further indication that .. .. ... 65.2 .. ... ... 414 .. ... ... ... 420 Si:2 ... oils can be identified as true plastics above ii:8 ... 59:s . . . . . . . . . . . 423 the limiting rates of shear, since in ac... .. .. ... 445 80'2 ... 75.5 .. 69 ... .. .. .. 454 .. ... cordance with expectation the intercept .. ... .. . . . ... 467 80.7 ... increases with an increase in top rate of .. 64:3 ... .. .. .. .. ... 474 .. .. .. .. shear, or top r. p. m., as shown in Figure 6 .. 73 .. ... 495 81 . . . . . ai:5 . . . . . ... ... 520 8 2 . 3 and Table 11, A. Measurements also .. .. ... 75,9 .. ... .. ... 523 show that a complete recovery of structure .. .. ... . . . ... 344 81 9 takes place if the oils are left a t rest for .. .. .. ... ... ... .. 568 81.7 ... 71 .. ... ... 79 ... 569 a period of time following shear agitation. ... . . ... ... . . . 595 si:5 . . ... .. 73.8 ... ... .. ... 621 81 5 Since this is one of the most important ... characteristics of thixotropy, it bears out .. ... ... .. ... ... 648 81 ... i6:3 .. ... .. ... ... 668 the statement that oils show thivotropic .. .. , , . .. ... . . . 676 80:5 ... .. ... ,.. .. ... ... 700 79.3 behavior above their limiting rates of shear. t . .
I..
428
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 15, No. 7
TABLEIV. HEAVY MINERAL OIL TORQUE-TIME CURVES AT VARIOES CONSTANT R. P. &I. (FIGURE 8) Time
(Temperature constant, 30' C.) Torque X 10-6 for Various Thixotropic Levels Torque X 10-6 for Various Thixotropic Levels 73 r. p. m. 140 r. p. m. 200 r. p. m. 300 r . p . m. 400 r. p. m. Time 73 r. p. m. 140 r. p. m. 200 r. p. m. 300 r. p. m. 400 r. p. m.
Sa.
Sec.
0 3 4 7 8
2.67
..
4: 35
11 13 14 15 18 19 20 23 24 25 28 29 30 34 35 40
2159
..
3:92
2:58
3:s3
in
50
60 70 80 90
100 110 120
2:61 2.6
..
2158
.. 2:57 ..
4.55 4.54
..
..
3:s 3:77
2156
.. ....
2156 2.55 2.51 2.51 2.5 2.49 2.48 2.47 2.45 2.44
3:fl 3.68 3.65 3.61 3.57 3.54 3.5 3.47 3.43 3.4
3175
6.09 5.12 4:81 4:67
..
4:56
..
4:49
..
4:39
..
4:39 4:35 4:32 4.22 4.16 4.1 4.04 4 3.95 3.91 3.89
8.23 7.09
..
6:9
..
6183
10.6
...
8.5
8:i4
.. .. 7:s .. ..
.. 6177 .. 6172 .. .. ..
7:i
6:67
7:03
6:58 6.53 6.44 6.33 6.23 6.12 6.04 5.98 5.91 5.82
7:68 7.51 7.38 7.21 7.1 6.95 6.81 6.71 6.61 6.51
..
7168
..
Experimental Curves A iarge number of oils have been measured at various top r. p. m. Since all are of a similar nature, only one representative flow curve is shown in Figure 7 and Table 111. Figure 8 and Table IV show the decrease in torque as a function of time for constant rates of shear or r. p. m. This decrease is typical for thixotropic plastics.
FIGURE 7. EXPERIMENTAL FLOWCURVES OBTAINED FROM VARIOUS TOPR. P.M.
130 140 150 160 170 180 190 200 210 220 230 240 270 300 330 360 390 420 450 480 510 540 570 600 630 660 690 720 750 780
2.43 2.41 2.4 2.39 2.38 2.37 2.36 2.35 2.35 2.34 2.33 2.32 2.3 2.28 2.26 2.28 2.24 2.21 2.2 2.2 2.2
..
.. .. *. .. .. .. ..
3.37 3.35 3.34 3.3 3126
.. ..
3:ia 3:i3 3.08 3.05 3.01 2.98 2.94 2.91 2.9 2.87 2.85 2.84 2.84 2.8 2.8 2.78 2.76 2.73 2.73 2.71
3.84 3.8 3.77 3.75 3.71 3.67 3.65 3.63 3.61 3.57 3.54 3.52 3.45 3.4 3.35 3.34 3.3 3.28 3.26 3.23 3.2 3.19 3.15 3.14 3.11 3.1 3.08 3.06 3.05 3.05
5.78 5.72 5.7 5.61 5.6 5.56 5.5 5.48 5.41 5.4 5.38 5.33 5.27 5.2 5.13 5.11 5.08 5.05 5.02 5.0 4.98 4.98 4.95 4.95 4.94 4.92 4.92 4.92 4.92 4.92
6.43 6.33 6.28 6.19 6.12 6.08 6.01 5.97 5.91 5.88 5.83 5.8 5.7 5.61 5.54 5.49 5.46 5.4 5.38 5.33 5.32 5.3 5.3 5.28 5.28 5.28 5.27 5.27 5.27 5.27
Temperature, Turbulence, and Slippage
It is evident that the type of flow curves obtained from oils is not caused by factors l i e change of temperature, turbulence, or slippage. Slippage can be immediately discarded as a result of experiments made with a grooved bob and cup (3). Turbulence can be eliminated as cause for the particular structure of the oil flow curves, since turbulence would have tended to increase rather than decrease the forces (torques) at higher rates of shear. Finally there remains the question of temperature. Increase of temperature undoubtedly decreases viscosity. The point then is, how much rise in temperatwe is required to decrease the oil viscosity to the same plastic viscosity obtained by increasing the rate of shear. To determine this, a heavy mineral oil was chosen and its viscosity below the limiting rate of shear %as measured a t various temperatures. Though the decrease in viscosity is substantial, it is not large enough to account for the rapid decrease in viscosity with an increase in top rate of shear. I n Table I1 the true Newtonian viscosities are given for various temperatures and the plastic viscosities are given for various thixotropic levels and are designated by their respective top rates of shear and top r. p. m. Table I1 makes it obvious that the decrease in plastic viscosity resulting from an increase in top rate of shear is too great to be entirely caused by an increase in temperature. This point is more clearly shown by plotting the plastic viscosities at various top r. p. m. against the true Newtonian viscosities obtained a t various temperatures (Figure 9 and Table V). For example, an increase from 100 to 700 top r. p. m. decreases the plastic viscosity from 710 to 150 poises, which in turn requires a temperature increase from 31 to 48" C., if the temperature is responsible for the entire increase in viscosity. However, no appreciable temperature increase could be observed if the temperature was taken before and immediately after the measurement. Therefore temperature can also be ruled out as a determining factor for the particular shape of the oil flow curves.
429
ANALYTICAL EDITION
July 15, 1943
HEAVY MINERAL OIL
l a ) ( b ) ( c ) S4MPLE I (d) ( e )
70(
S4MPLEZ
R P M :CONSTANT
60
501
z
a
a a 40( 0 I-
30(
0
20(
T I M E IN SECONDS
CCRVESAT VARIOVSCONSTAKT R. P. M. FIGCRE 8. TORQUE-TIME
TABLEV.
.
101
INTERPOLATED DATAFROM FIGURE 9
Viscosity Poiees
Top R. P. M.
Temperature
660 600 500 400 300 200 150
110
31 32 34 36.5 39.5 44.5 48
155 225 305 415 580 700
c.
Conclusions A full theoretical treatment of the phenomenon of the thixotropic behavior of oils has not yet been developed, but a few suggestions may elucidate the results so far ,presented. At rest all molecules are distributed a t random, taking on a statistic average position. Applying a rate of shear beyond the limiting rate of shear, which is equivalent to developing a minimum directional force, the molecules may start an alignment in the direction of shear; hence the original random structure of the oil breaks down, as indicated by a decrease in plastic viscosity. Upon rest the molecules will slowly return to their original random position. Discussion According to the literature, oils are often recommended for use in calibrating viscometers, but this may lead to a serious error (3). The present paper shows that unreliable results may be obtained if oils are used above their limiting rates of shear, since most oils behave like true Newtonians only below this critical point. The danger of an unreliable calibration with oils is particularly great if highly viscous oils are used, since their limiting rates of shear are very low. However, any such oil can be used for calibration if the applied rates of shear are kept below the limiting rate of shear of the oil, in the range where it behaves like a true Newtonian liquid. Although the oils tested showed no evidence of impurities, some oils contain waxes and other contaminants; therefore, the question of a possible separation at high rates of shear remains to be discussed. Separation would probably have a decreasing effect on the torque and therefore would show
TEMPERATURE
FIGURE9.
pc.)
CURVNCONSTRUCTED
TO SHOWINCREASE IN TEMPERATURE REQUIRED TO OBTAIN SAME \’ISCOSITY AS OBTAINED BY INCREASE IN T O P R. P.M.
phenomena like those due to a thixotropic breakdown of the oil structure. However, the effects described above are not due to any detectable separation, since repeated measurements after short periods of time yield identical results, and the elapsed time intervals, of a few minutes, are long enough to allow a thixotropic recovery, but not a redistribution of any separated materials. It is highly improbable that a redistribution or redispersion could take place while the material is at complete rest, even if a longer time for recovery were allowed.
Acknowledgment The author is indebted to Interchemical Corporation for permission to publish this work, to Henry Green for his valuable advice and suggestions, and to Evelyn Berezin for assistance in producing experimental data. Literature Cited (1) Buckingham, E., Proc. Am. Sac. Testing Matwiala, 21, 1164 (1921). (2) Freundlich. H., “Thixotropy”, Paris, Hermann & Cie., 1935. (3) Green, Henry, IND. ENG.CHEM.,ANAL.ED.,14, 676 (1942). (4) Green, Henry, Proc. A m . SOC.Testing Materials, 20, Part I1 451-94 (1920). (5) Green, Henry, and Weltmann, R. N., IND. ENQ.CHEW,ANAL. ED.,15, 201 (1943). (6) HatsohBk, E.,Kolloid Z . , 13, 88 (1913). (7) Houser, E. A., J. Rheology, 2, 5 (1931). (8) Ostwald, Wo.,and Stuart, W. W., Kolloid Z . , 78,324 (1937). (9) Peterfi, T.,Arch. Entwicklungsmech. Organ., 112, 660 (1927). (10) Pryce-Jones, J., J . Oil Colour Chem. Assoc., 17, 305 (1934). (11) I b i d . , 19, 293 (1936). (12) Reiner, M., and Riwlin, R., Kolloid Z . , 43, 1 (1927). (13) Schalek, E.,and Saegvari, A., I b i d . , 33, 326 (1923).
